Re: N-pt DFT where n != power of 2

dspdummy wrote:
On 2009-08-13 12:53:55 -0300, Jerry Avins <jya@xxxxxxxx> said:

Jerry Avins wrote:

...

zeroes to bring the larger up to 1024 and the smaller to 120.
128!

Jerry
120 = 3 * 5 * 8 and 960 = 8 * 120 = 3 * 5 * 64 so both are in the
small primes of 2, 3 or 5 range.





Can't pad zeros and do the 1024 or 128 fft(it's really an ifft that I
need) as the output time samples must respect the 960 or 120 frame size.
The results have to be bit-exact with those of the 120 or 960 dft. I looked
at that Wikipedia page but looks pretty complicated. I'll try to have a
look at the other website with some useful programs (probably prime factor
fft as was mentioned above). If there are any Matlab m files emulating this
I would be grateful to get my hands on them as it's quite helpful for
understanding the theory (I mostly speak C,Matlab,and block diagrams, and
not too much formulas)

Zero padding doesn't change the returned values. The extra zeros are place holders and can be removed from the final result, which are otherwise bit exact. You should do a little reading (or use FFTW, reading the directions carefully and taking no steps for granted.

Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
.

Relevant Pages

  • Re: What Are You Reading??
    ... from it reading other things; just finished Marcus du Sautoy's "The ... Music of the Primes" (popular history of mathematics). ...
    (rec.pets.cats.anecdotes)
  • Re: read/write at the same time
    ... to it in order to create other primes. ... procedure allows me reading from it and rewrite allows me only writing ... sequence of fixed-length records, where you can change one record ...
    (alt.comp.lang.borland-delphi)
  • Re: Methods that count primes without counting primes or referring to them...
    ... > I've been reading over about as much stuff as I can find about prime distribution and prime counting, and I've been left with the following question. ... function for generating the number of primes less than n that doesn't ...
    (sci.math)
  • Re: Methods that count primes without counting primes or referring to them...
    ... > prime distribution and prime counting, ... > Assuming I'm understanding what I'm reading correctly (I'm getting ... > Riemann's paper on the distribution of primes provides a function for ...
    (sci.math)
  • a 50 $ question on primes
    ... Tried to prove this for two full months and failed (a friend verified it on a coputer for triples of primes up to about 12000). ... I'll send the 50$ check to anyone who can prove/disprove it (to be proudly put on the wall in a nice frame). ...
    (sci.math)